GraySoft
Projects Models Compare Cloud benchmarks FAQ Download guIDE →
Model Intelligence Sheet

WariHima/hourai3-yomi-90m-v1-gguf overview

hourai3 yomi 90m v1 zenz v3互換 読み推定モデル architecture: lfm2moe tarined by: rtx3060 8m training code in ./training codes tokenizer from ku nlp/gpt2 small japanese …

ggufjabase_model:WariHima/hourai3-yomi-90m-v1base_model:quantized:WariHima/hourai3-yomi-90m-v1license:cc-by-sa-4.0endpoints_compatibleregion:us

Runs locally from ~162.5 MB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).

Downloads
0
Likes
0
Pipeline
Author

Repository Files & Downloads

1 GGUF files detected
Direct downloads for local inference
FileTypeQuantizationSizeLink
hourai3-yomi-90m-v1.ggufGGUFGGUF162.5 MBDownload

Model Details

Model IDWariHima/hourai3-yomi-90m-v1-gguf
AuthorWariHima
Pipeline
Licensecc-by-sa-4.0
Base modelWariHima/hourai3-yomi-90m-v1
Last modified2026-08-02T13:59:22.000Z

Model README

---

license: cc-by-sa-4.0

language:

  • ja

base_model:

  • WariHima/hourai3-yomi-90m-v1

---

hourai3 yomi 90m v1

zenz v3互換 読み推定モデル

architecture:

  • lfm2moe

tarined by:

  • rtx3060 * 8m

training code in ./training_codes

tokenizer from ku-nlp/gpt2-small-japanese-char

+(unused qwen3 special toknes from base model)

dataset:

  • Miwa-Keita/zenz-v2.5-dataset

use wikipedia part 40k rows 1 epoch

qunat type

  • BF16

example code (python)

import llama_cpp

llm =  llama_cpp.Llama(
     model_path="./hourai3-yomi-90m-v1.gguf",",
        embedding=False,
        verbose=False
    )

while True:
      input_text = input("入力した文章の読みが出力されます。exitを入力で終了:")
    
    if input_text == "exit":
        break
        
    text = f"<s>\uEE00{input_text}\uEE01"
    
    output_text = llm(
        text,
        max_tokens=10,          # 生成する最大トークン数
        temperature=0.0,       # ランダム性(0.0に近づくほど確実な出力、1.0以上で多様化)
        top_k=40,               # 上位k個の候補に絞り込む
        top_p=0.95,             # 累積確率p以下の候補に絞り込む
        repeat_penalty=1.1,     # 同じ単語の繰り返しを抑制するペナルティ
        stop=[],
        echo=False
    )["choices"][0]["text"]
    print(f"入力|出力: {input_text}|{output_text}")

convert

変換時、qwen3nextがgpt-2トークナイザで使用することを想定されていなかったため、
llama.cppリポジトリのconversion/base.pyファイルの
以下のエラーをバイパスする必要がありました。
1687=1699付近
        if res is None:
            logger.warning("\n")
            logger.warning("**************************************************************************************")
            logger.warning("** WARNING: The BPE pre-tokenizer was not recognized!")
            logger.warning("**          There are 2 possible reasons for this:")
            logger.warning("**          - the model has not been added to convert_hf_to_gguf_update.py yet")
            logger.warning("**          - the pre-tokenization config has changed upstream")
            logger.warning("**          Check your model files and convert_hf_to_gguf_update.py and update them accordingly.")
            logger.warning("** ref:     https://github.com/ggml-org/llama.cpp/pull/6920")
            logger.warning("**")
            logger.warning(f"** chkhsh:  {chkhsh}")
            logger.warning("**************************************************************************************")
            logger.warning("\n")
            
            return "default"
            #raise NotImplementedError("BPE pre-tokenizer was not recognized - update get_vocab_base_pre()")
想定されていないことによる変換時のエラーなので、推論時はmainstreamのllama.cppで動作します。

Run WariHima/hourai3-yomi-90m-v1-gguf with guIDE

Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.

Download guIDE → · Browse 524k+ models · Compare models

Source: Hugging Face · Compare models